Population Model v. 0.1

[The Joker]

It is great that you are posting your ideas here, btw. Thanks!

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website

[DarkCloud]

The way I decided my groups was I looked at how the ages are in their relationship to each other.

42-49 are the rapidly aging middle class that are sort of exiled from computer work, but don't quite want to retire until 50-60. In past ages this age group was the oldest that it could be expected to live to.

quote: It is great that you are posting your ideas here, btw. Thanks!

Well you are welcome; I stop by here from time to time; I also posted in the Economy Thread 0.1 (I think)

[The Joker]

I think culturally decided age groups should be totally avoided. Maybe in some other culture people would retire or begin working at other ages.

The age group system should be as objective as possible, and that is done best with 10 year groups. Then the cultural things will effect at what ages people begin to work, retire or die. And via the age groups the amount of people in each category could be found easily.

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website

[The Joker]

To comment on Chris' pop model program I think that it is really great. It is just what I had in mind. So great work, Chris!

It should be possible to use the things from that program in a more complex economy and pop program, where food per pop unit effects the mortalities.

Just one question: How do you decide the birth rate in the program? I went through the source code, and I couldn't find it. Is it just a constant?

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website

[Sikander]

Well, I just finished reading all of the posts here, whew! Here are a few random thoughts:

Warfare: Historically warfare itself had very little effect on population directly. In Hunter-Gatherer societies the population base and the concentration of populations were so small that warfare was usually ritual, with very few casualties.

With advanced agriculture (ie agriculture including grains, legumes and domestic animal proteins which can actually support life in the long run in a small area) came specialization and professional troops. While the population increased and the tempo and intensity of warfare increased with it, the low surpluses of early agriculture meant that the militarized percentage of the population remained small. Thus even a brutal casualty rate amongst the soldiery had very little effect upon the size of the population as a whole. Another factor to keep in mind is that the fertility rate is much more sensitive to the population of fertile females than it is to the population of fertile males. This would of course be more the case in a society which allowed multiple wives for men, or for mistresses to be kept, but is nonetheless true even in societies with fairly strict monogamy.

Warfare does have some effect upon population even outside of the fairly rare war of annihilation. This effect usually has two faces. In the first case, returning soldiers may bring home diseases which can ravage the weaker children and elderly especially. The other case involves stresses to the civilian population in the war zone. This can be caused accidently as civilians flee their homes and leave crops untended, and the weak are brought low by lack of shelter and deprivation, or purposefully as a policy of terror (particularly popular with the Turkic peoples like the Mongols, Huns etc.) to drive the population from the land through terror and physical devestation of the infrastructure. Siege warfare is another special case, where an urban population is cut off from it's lifeline to starve or surrender.

The wars of the past couple of centuries are the exception rather than the rule. I am talking about the World Wars, the Napoleonic wars and the American Civil War here, where huge citizen armies fought unrestricted total wars for enormous political goals (ie to completely vanquish their enemies). The norm throughout history is a limited war fought by small (usually professional) armies for a very limited aim (a border province, a city, loot).

What does this have to do with the population model? Well quite a bit actually.

1) The fertility index should be more sensitive to the number of fertile females than males. A fairly simple way to do this would be to calculate the fertility base as (The number of fertile females) - (X / Y) where X equals the difference between the number of males and females (assuming fewer males) and Y is a cultural factor where a low number assumes polygamy and a high number assumes strict monogamy.

2) Warfare on enemy turf should have very little effect upon the home front, with the exception being that disease may find a way home along with any troops who return.

3) Warfare on home turf may have a huge effect, as stressed populations may be forced from their lands (lowering production) and become refugees (and perfect disease vectors) fleeing into adjacent provinces. The effect of this will vary somewhat depending on the attacking force's culture and aims. Raiders will tend to use terror more than forces pursuing a persisting strategy (ie conquest), and nomads will also be more likely to devestate the land to create a low population area more consistent with their needs. Thus an agricultural neighbor with similar cultural traits bent upon permanent conquest of a border region may cause very little emigration from the invaded province, while Tamerlane could create a wave of panic and destruction behind your lines.

[S. Kroeze]

Dear Amjayee, Joker and others,

I would gladly help, though generally I am quite busy, so please don't expect too much. It will probably take some days before I will leave the house to visit the library. Yet I promise to try at least to do something.

Where can I find the demo? I read about it on the Forum but couldn't find it. By the way, when it's all in programming language I wouldn't need it, because I have no bent for the exact sciences, though I have some understanding of statistics.

For the moment the only advice I can think of is to search for one -or preferably several- of the following reliable studies:
D.V.Glass and C.Eversley: 'Population in History: Essays in Historical Demography'(1965)
E.A.Wrigley: 'Population and History'(1969)
C.M.Cipolla: 'Economic History of World Population'(1974)
T.McKeown: 'The Modern Rise of Population'(1977)
A.M.Carr-Saunders: 'World Population:Past Growth and Future Trends'(1936)
E.A.Wrigley and R.Schofield: 'The Population History of England, 1541-1871'(1981)
'United Nations Demographic Yearbook'
T.Chandler and G.Fox: '3000 Years of Urban Growth'(1974)
P.Bairock: 'Cities and Economic Development'(1988)
P.M.Hohenberg and L.H.Lees: 'The Making of Urban Europe, 1000-1950'(1985)
B.R.Mitchell: 'International Historical Statistics,1750-1988:Europe(1992)
W.H.McNeill: 'Plagues and Peoples'(1976)
W.H.McNeill: 'The Global Condition: Conquerors, Catastrophes, and Community'(1993)
T.McKeown: 'The Origins of Human Disease'(1988)
K.F.Kiple(ed.): 'The Cambridge World History of Human Disease'(1992)

N.B.: I do not know which of these titles will prove to be most useful. Some of them will doubtless refer to other literature. I am sure a well-assorted university library will have at least most of these titles.

By the way, your question for information about death rates of different age groups is more or less unanswerable, because many people weren't well-fed and there was almost always some contagious disease around. And hygienic conditions were generally bad or exceptionally bad, especially in the cities. Another problem are our limited data for the more ancient periods. Most population figures for these periods are just intelligent guesses, nothing more!

So you have to make the most of all information you can find. Also very useful are present-day data dealing with some developing country like India or Nigeria. I think when you would use all figures assembled by me, you have already something to start with. When you have figures for life expectancy at birth, average birth and death rates, infant mortality and some idea about the structure of the population pyramid, you can make a quite intelligent guess answering your question. My advice is to make a guess and to play around with your figures to see whether the result seems realistic. Afterwards you can control whether the development of the population after a century or more is in accordance with those variables you knew beforehand, like average life expectancy etc. Casualties of war are generally too few to influence events that much, though there are notable exceptions, like Cannae.

And to reassure you: I honestly think that as long as you use plain common sense you cannot err too gravely. This is only a first model, there will remain an opportunity to correct some details!

[Sikander]

Ethnicity, Language, Culture, Religion, Nationality:

This is a very complex set of circumstances. I can see why Civ opted to ignore these factors completely, yet by doing so they cut the heart out of the game. Perhaps a good starting point would include some definitions:

Ethnicity - from Greek meaning a people or a nation. For our purposes we can probably state that Culture + Language = Ethnicity.

Language - should generally be understood.

Culture - can be summarized as the values and traditions of a group. For our purposes this should probably not include religious values, which should be handled seperately.

Religion - a system of belief to explain and guide people through the portions of human existence which are not otherwise easily knowable. This should have it's own submodule, as it can obviously be very different from one Religion to the next. Along with Culture is the most important determinant of societal values.

Nationality - can be assumed to basically be ethnicity. When aroused by Nationalism, will tend to create a demand for a State based upon ethnic identity. This can have dire consequences for empires.

-----------------------------

I have noticed that a lot of discussion has been held regarding the tendency for groups to vary over time, and eventually become completely seperate groups with their own ethnic identity. While this is a valid and observable phenomenon, IMO it is in fact the opposite which is the dominant trend of recorded history. Do not be fooled by the current post-colonial and post cold war trend of irredentism. The fact is that since the dawn of history the number of languages, cultures peoples and states has been coming together (becoming fewer). This trend will continue in the long run, especially as communications improve. We are after all a group of people who speak many different languages at home who are working together over the internet in English on a project.

The tendency to pull together is stronger in some areas of human endeavor than in others. It is probably strongest in the area of technology, where Iranian Mullahs are just as happy to get Tanks and Missles as Israeli Rabbis are. They may well decry the evil of the societies which invented and perfected those weapons, and be very closed to other ideas coming from those societies, but the technologies themselves seem to cross borders free from judgements of this kind.

Language is another cross border traveller, though it does not travel as easily as technology, perhaps because it is part of what defines an ethnos. Nonetheless language does cross borders, whether in the form of a few words (like the word Tank for example) which travel along with a new concept, or in the form of a lingua franca which can be used far and wide amongst people of various languages for trade, diplomacy and exchanges of ideas.

Interestingly, language does not behave in the manner often ascribed to it, where the further from the center of it's geographic center the more changed the language. It is in fact the opposite, that is to say that language changes fastest at the center, and these changes propogate to the periphery. Basically language changes at a rate based upon the number of speakers. Thus Latin changes very slowly, while English changes very quickly. Populations which are small and isolated change hardly at all, which is why Icelandic is very much closer to the Norwegian spoken centuries ago than modern Norwegian is today.

Religion is another concept which can spread quickly throughout a region without respect of borders. A model of religion might in fact have a seperate map superimposed over the political map where religions do battle over populations much as empires do. In many respects this is a fairly modern phenomenon. Religions seem to have developed like States and Warfare, from a very local and personal entity which was capable of cohabitation into global conglomerates which in some cases seek to encompass the entire planet and drive out all others.

---------------------

Why is any of this important to the Population Model in GGS?

Because people are often subsumed by Cultural, Technological, Linguistic and Religious factors with little or no action taken by political forces other than to react to these changes. Thus Greek and Roman traders bringing strange goods from afar offered many more things to a local tribe than the contents of their holds. They brought Laws, Coins, Cultural ideas, a language which allowed communications throughout a vast area, and numerous technologies. It was in this way that many tribes became Romanized, ready and sometimes willing to be fit into the fabric of the empire.

Conversely this Empire which straddled so much of the Western world was itself conquered by Christianity despite it's efforts to resist. The lesson is that strong ideas push out weak ideas. Boudaries of Religion, State force, Culture and Language can slow this process, but are rarely capable of stemming the tide forever.

[The Joker]

Sikander:

Warfare:
Well, yeah. But it still has to be implemented. Even in ancient wars meant that many people were away from their land, which could cause food shortages and famines. So creating armies would take up a considerable amount of people. Many of them would return home, yes. But if they are away for years and years it would have some effect.

Polygamy:
Maybe this will add too much complexity. It wouldnt really have any huge importance.

Otherwise I agree. Great thoughts you have made.

Your second post:

Ethnicity etc:
I think to make it simple that language, ethnicity and nationality the same: nationality. Nationalities would have a number showing how similar they are to one another. If it is low then they are very similar, if it is high then they are not very similar. Religion would be handled seperately, with the religious leaders having effect on their believers. Culture I am not sure how to handle. Granted, it should propably have some effect on different thing. But what and how?

Assimilation:
Yes, I agree that the long run trend would be towards assimilation. But on the other hand it should be very hard to suddenly assimilate a lot of people. The Roman Empire did very poorly here, and was suddenly broken into loads of small civs. So colonies have to develop their own nationality. That is required for the rise and fall idea to be implemented.

I pretty much agree with all this. It is a very hard thing to implement in a game, and therefore it will require a lot of consideration from us, before we can make it work. These are definately some good thoughs.


S. Kroeze:

Thanks again for your help.

Demo:
Right now we are working on a more sophisticated UI to go with the demo, so it will be really playable. Right now we dont have any real demo. A few prototypes in different areas, but nothing playable. VetLegion made a textbased political demo/game, that is a bit playable. I could mail that to you if you want it. It doesn't have that thrilling gameplay yet, though.

Death Rates:
The thing is that we need some way to have a "basic" moretality, and then poor nutrition should increase this. The same with diseases.

But I hope you are right that with the right infant mortality rates and the life expectancy we couldnt go too wrong. I guess we will just have to play around with it, then.

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website

[Leland]

While tryign to chew this population model, the following thougth experiment came to my mind:

A player rules a smallish civ with only one region. However, the scarcely resourced northern tundra is obstructing the population of even the southmost cities from reaching an age structure which would allow concentration on educating the people. So, the player decides to divide his empire into two regions: the underdeveloped north and a few small successful cities from the south. Southern region would have more resources, and with proper social engineering it could use those resources for changing the age structure (or whatever property) to the civs favour.

The problem I see with this is that the player is not doing anything but changing the way the civilization's resources and population are calculated. Hmm. Maybe I didn't quite comprehend the model... what is a region exactly? Are there some limitations in regioning the civilization? Is it even up to the player? If not, how are regions defined?

[cavebear]

I see 2 opposing trends here. The first moves toward complexity (emmigration, age groups, fertility, etc). The second moves toward simplification (avoiding player micromanaging). I don't think you can do both.

It is one thing to incorporate population changes according to age groups, etc, but it is quite another to make those age groups *relate* to player actions and other game factors.

The first idea broadens player options in controlling the play of the game (and many of them would be greatly appreciated by micromanagers). But the second is contrary to that. Several of you have mentioned that the human player will have only limited control over broad outcomes and that many functions will play out beyond player control entirely.

I suggest that you be cautious of frustrating a player too much by allowing a civ to operate too independently. As a personal example, I can assure you that if I get too frustrated by seeing my civ go in directions I cannot manage at least indirectly, I'm going to stop playing the game very quickly.

In other words, If I can control taxes, but not where the money goes (tech, food, religion, military, etc), I have no reason to play. It would be more like watching a computer-generated movie than playing a game.

Does that make sense?

[The Joker]

Leland:

No, I think having a scarcely ressourced northern tundra would just make people move away from it. This would mean you would have a region where the vast majority of the population lives in the south. So it shouldn't become a burden to your civ to have an area without that many ressources, and dividing your civ into two regions shouldn't do any difference either.


Cavebear:

Although that makes some sence I think that it could be overcome by having the player be only the leader of the government. He would decide how much the tax levels would be, and what to use his money for. And he could do things to effect economy. But he couldn't control it. Population age groups are mostly just an effect of technology levels and food per capita (all depending on the overall shape of the civ, but not on small civ characteristics). Age groups is not only a cool spice to give to the game, it also makes some cool possibilities possible, like going to war would kill many young males, leaving a lack of labour or something. But it wouldn't be anything the player would have to micromanage, and most likely he wouldn't be able to.

------------------
"Life is a lesson. You learn it when you're through."
- Limp Bizkit

GGS Website

[Leland]

quote: Originally posted by The Joker on 12-30-2000 08:06 AM No, I think having a scarcely ressourced northern tundra would just make people move away from it. This would mean you would have a region where the vast majority of the population lives in the south. So it shouldn't become a burden to your civ to have an area without that many ressources, and dividing your civ into two regions shouldn't do any difference either.

Okay, but I still can't quite grasp the concept of a region... has anyone come up with a valid definition? the one the the website talked about regions being an aggregate of cities, but in this thread you've been talking about regions being the basic units in a civ. Hmm.

Now that I've reconsidered my example, I still tend to think that you could tweak certain characteristics of populations by merely splitting them up or combining them in specific ways. You could prevent or endorse differentiation in your overall population by just drawing borders on a map. I can't see how migration would change this. To stick with my previous example, let's assume that the population has already settled to an equilibirium (that is, the vast northern areas are scarcely populated in contrast to the dense but few southern cities) and that the south produces more trade than the north. If the player splits the civ to two regions, the north would suddenly not get the benefits of the trade (this could happen indirectly, for instance if the land just can't support a large merchant class) whilst the south would no longer need as many farmers (because infertile land of the north is more difficult to cultivate). The populations would differentiate, yes?

I must have a blind spot here somewhere. Please point it out.

As for micromanagement... even if the game would not allow it, who says the internal game mechanics shouldn't go that deep? Jumping from cities to regions is kind of a step back to a more simplistic viewpoint. I'd like to see regions as some sort of emergent constructs to tackle the ever increasing complexity.

(Just explain what regions are and I *may* change my mind, of course.:P)

[amjayee]

About avoiding micromanagement, the idea was, that most things can be tweaked with; it's just that those options are usually handled by the AI, and the player can, but doesn't have to touch them; and after all dealing with everything would become quite tedious.

About player not having enough control over his empire, that will not be so... the player can choose the level of involvement he gets, by choosing his government style. And let's remember this is not civ, the game objectives will be different and will not require total control over all things. I'm sure it will be fun, it's just not easy to see it currently: and to be honest we are not helping it, by letting the design discussion sprawl around like this. Getting the design doc is a really great thing.

Well, when the UI and map are done, I'm sure everything starts to get easier. And in the alpha stage, we have overcome the biggest obstacles. Just stick with us and help us clear these things out, since we are still only mid-way to understanding the game system properly.

[Leland]

I agree 100% on the micromanagement issue; I was just a little bit confused what you were after.

But on the population mode I still have my doubts. It should not be region based if regions are defined by the player, that's just not realistic. If you have a multicultural nation and you just wipe off your province borders, the population doesn't magically become uniform. This is why I suggest that you use your original idea of "population units", but with the exception that they span over several tiles. Of course an array of these tiles will have to be stored and the memory requirement grows geometrically with the size of the population... but let's not totally bash the idea.


Let's estimate a civ with about 1000 different populations (quite possible). The characteristics (age structure etc.) of a population migth take up to 1 kilobyte (at least they won't take more?). Most urban populations, and thus most of the populations, since cities tend to be "melting pots", span only across one or a few tiles. The largest uniform populations might span across the area of tens of thousands of tiles, but their number would be smaller. So, let's assume an average population size 1000 tiles.

Thus

mem_requirements = number_of_pops * (sizeof_popdata + average_pop_area * sizeof_coordinates) = 1000 * (1k + 1000 * 4 bytes) = 5Mbytes.

Quite a lot. Is it doable? Can the number of populations be reduced somehow without resorting to the predefined regions? Can more efficient compressing schemes be used for storing the area populations are in? (I noticed I entirely forgot the number of people per each tile... it's going to cause an additional 2Mbytes or so to be used... darn.)

Population model seems to be one of, if not the most important model. Everything revolves around people. That's why I think the model should not be coupled with the regional/governmental models which in my opinion are higher level concepts. Also, now that I started rambling I migth as well say that the populations should *inherently* be capable of being mobile populations; i.e. there would be no difference between a mobile population and a "normal" population. This coudl maybe be achieved by letting the "controller" of a population to be a unit in addition to a city or a region?
[This message has been edited by Leland (edited January 02, 2001).]

[cavebear]

Regions? Well, I think of a "region" as a geographically cohesive area. In the US, we have the Northeast (cold winters, cool summers), the Mid-Atlantic (cool winters, warm summers), the South (warm winters, hot summers), the Midwest (dry and hot), and West Coast (warm, temerate, and wet).

The geography of a "region" defines the lifestyles and culture to a significant degree. I think that 2 or more civs might split a natural terrain area, and each part might be a "region" for each civ, but I don't think that desert and jungle should be combined by player demand into a single region.